Predictors of positive stool culture in adult patients with acute infectious diarrhoea

Abstract

Introduction: Stool cultures for bacterial pathogens are often requested for investigation of patients with infectious diarrhoea, but the literature reports low yield. This study aimed to identify the predictors of positive stool culture in adult patients presenting to the Emergency Department (ED) with acute infectious diarrhoea, and to assess the relative importance of these predictors. Methods: A retrospective case-control study was performed in the setting of an ED of an university-affiliated hospital in Hong Kong. The inclusion criteria were: a. Adults (age >16). b. Presentation to the ED with acute diarrhoea (of duration not longer than 2 weeks). c. Outpatient treatment with or without a brief observation unit admission. d. Stool cultures requested from the ED. Patients were excluded if there was an obvious non-infectious cause of the diarrhoea at presentation, such as Crohn’s disease, radiation colitis or irritable bowel syndrome. We compared 130 consecutive cases with positive stool culture, over a 12-month period, against 119 control cases obtained from a random sampling of 524 consecutive negative cases over the same period. Univariate analyses (chi-square, Mann Whitney U and t-tests) and multivariate analyses were performed to compare variables. Results: By stepwise logistic regression analysis, factors found to be independently associated with positive stool culture were: body temperature (fever) 37.5°C - 38.4°C (OR 3.49; 95%CI 1.44-8.48); duration of abdominal pain longer than 3 days (OR 0.27; 95%CI 0.12-0.64) - negative correlation; requirement of intravenous fluids (OR 2.35; 95%CI 1.13-4.92); and month of presentation (season) from May to October (OR 3.90; 95%CI 1.93-7.86). Bloody diarrhoea was not predictive (OR 1.19: 95%CI 0.52-2.72). Conclusion: Our study identified several important independent variables to be predictive of positive stool culture in the adult ED setting. Further studies are required to derive a prediction rule, for more cost-effective utilisation of stool culture testing.